├── .github
    └── workflows
    │   └── test.yml
├── .gitignore
├── .gitmodules
├── CHANGELOG.md
├── Makefile
├── README.md
└── simhash_server.cpp


/.github/workflows/test.yml:
--------------------------------------------------------------------------------
 1 | name: Test
 2 | 
 3 | on:
 4 |   push:
 5 |   pull_request:
 6 | 
 7 | jobs:
 8 |   build:
 9 |     runs-on: ubuntu-latest
10 | 
11 |     steps:
12 |       - uses: actions/checkout@v4
13 |         with:
14 |           submodules: recursive
15 | 
16 |       - name: make
17 |         run: make
18 |       - name: Start server
19 |         run: ./simhash.server &
20 |       - name: Wait for server to start
21 |         run: |
22 |           timeout=10
23 |           while ! curl "http://127.0.0.1:11201/?s=你好世界"; do
24 |             if [ $timeout -le 0 ]; then
25 |               echo "Server failed to start"
26 |               exit 1
27 |             fi
28 |             timeout=$(($timeout - 1))
29 |             echo 'sleep 1'
30 |             sleep 1
31 |           done
32 | 


--------------------------------------------------------------------------------
/.gitignore:
--------------------------------------------------------------------------------
1 | simhash.server
2 | *.swp
3 | 


--------------------------------------------------------------------------------
/.gitmodules:
--------------------------------------------------------------------------------
1 | [submodule "submodules/simhash"]
2 | 	path = submodules/simhash
3 | 	url = https://github.com/yanyiwu/simhash
4 | [submodule "submodules/husky"]
5 | 	path = submodules/husky
6 | 	url = https://github.com/yanyiwu/husky.git
7 | 


--------------------------------------------------------------------------------
/CHANGELOG.md:
--------------------------------------------------------------------------------
 1 | # CHANGELOG
 2 | 
 3 | ## v0.2.0
 4 | 
 5 | + Add git submodule for `submodules/` directory
 6 | 
 7 | ## v0.1.0
 8 | 
 9 | + Initial release
10 | 


--------------------------------------------------------------------------------
/Makefile:
--------------------------------------------------------------------------------
1 | all: 
2 | 	g++ -o simhash.server -O3 -DLOGGER_LEVEL=LL_INFO -Wall -g -I./submodules/husky/include -I./submodules/simhash/include -I./submodules/simhash/submodules/cppjieba/include/ -I./submodules/simhash/submodules/cppjieba/deps/limonp/include/ simhash_server.cpp -lpthread
3 | osx:
4 | 	g++ -o simhash.server -O3 -DLOGGER_LEVEL=LL_INFO -Wall -g -std=c++11 -I./include -I./include/CppJieba/ simhash_server.cpp -lpthread
5 | clean: 
6 | 	rm -f *.o *.a simhash.server
7 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # Simhash Server
 2 | 
 3 | [![Test](https://github.com/yanyiwu/simhash_server/actions/workflows/test.yml/badge.svg)](https://github.com/yanyiwu/simhash_server/actions/workflows/test.yml)
 4 | 
 5 | ## Introduction
 6 | 
 7 | HTTP Server based on [simhash]
 8 | 
 9 | ## Usage
10 | 
11 | Dependencies:
12 | 
13 | + `g++ (version >= 4.1 recommended) or clang++`;
14 | 
15 | ```sh
16 | git clone --recurse-submodules https://github.com/yanyiwu/simhash_server.git
17 | ```
18 | 
19 | ```sh
20 | make
21 | cd keyword_server
22 | ```
23 | 
24 | if you are using MacOSX, please use `make osx` .
25 | 
26 | ```
27 | ./simhash.server
28 | ```
29 | 
30 | ```
31 | curl "http://127.0.0.1:11201/?s=你好世界"
32 | ```
33 | 
34 | ```
35 | curl -d "我是蓝翔 技工拖拉机学院手扶拖拉机专业的。" "http://127.0.0.1:11201"
36 | ```
37 | 
38 | [simhash]:http://github.com/yanyiwu/simhash
39 | 


--------------------------------------------------------------------------------
/simhash_server.cpp:
--------------------------------------------------------------------------------
 1 | #include <unistd.h>
 2 | #include <algorithm>
 3 | #include <string>
 4 | #include <ctype.h>
 5 | #include <string.h>
 6 | #include "thread_pool_server.h"
 7 | #include "simhash/Simhasher.hpp"
 8 | 
 9 | using namespace husky;
10 | using namespace simhash;
11 | 
12 | const size_t PORT = 11201;
13 | const size_t THREAD_NUMBER = 4;
14 | const char* const DICT_PATH = "./submodules/simhash/submodules/cppjieba/dict/jieba.dict.utf8";
15 | const char* const MODEL_PATH = "./submodules/simhash/submodules/cppjieba/dict/hmm_model.utf8";
16 | const char* const IDF_PATH = "./submodules/simhash/submodules/cppjieba/dict/idf.utf8";
17 | const char* const STOP_WORDS_PATH = "./submodules/simhash/submodules/cppjieba/dict/stop_words.utf8";
18 | const size_t TOP_N = 5;
19 | 
20 | class ReqHandler: public IRequestHandler
21 | {
22 |     public:
23 |         ReqHandler(const string& dictPath, const string& modelPath, const string& idfPath, const string& stopwordsPath): _simasher(dictPath, modelPath, idfPath, stopwordsPath){};
24 |         virtual ~ReqHandler(){};
25 |     public:
26 |         virtual bool DoGET(const HttpReqInfo& httpReq, string& strSnd) 
27 |         {
28 |             string s, tmp;
29 |             httpReq.GET("s", tmp);
30 |             URLDecode(tmp, s);
31 |             uint64_t u64;
32 |             bool ok = _simasher.make(s, TOP_N, u64);
33 |             strSnd << u64;
34 |             return ok;
35 |         }
36 |         virtual bool DoPOST(const HttpReqInfo& httpReq, string& strSnd) 
37 |         {
38 |             const string& s = httpReq.GetBody();
39 |             uint64_t u64;
40 |             bool ok = _simasher.make(s, TOP_N, u64);
41 |             strSnd << u64;
42 |             return ok;
43 |         }
44 |     private:
45 |         Simhasher _simasher;
46 | };
47 | 
48 | int main(int argc, char* argv[])
49 | {
50 |     ReqHandler reqHandler(DICT_PATH, MODEL_PATH, IDF_PATH, STOP_WORDS_PATH);
51 |     ThreadPoolServer server(THREAD_NUMBER, PORT, reqHandler);
52 |     server.Start();
53 |     return EXIT_SUCCESS;
54 | }
55 | 
56 | 


--------------------------------------------------------------------------------